This is a document to highlight the steps involved conducting self organizing map analysis with data from the MRTA Projects.

Data are restricted acces and are not included in the repository

For more infomration please email : x.quek@garvan.org.au

Dependencies

library(edgeR)
## Loading required package: limma
library(pheatmap)
library(cluster)
library(R6)
library(reshape2)
library(plyr)
library(kohonen)
## Loading required package: class
## Loading required package: MASS
library(ggplot2)
source('src/somHelper.R')

Step 1 - Read data ( FPKM Value )

Step 2 - Working on Snail / on / off model only

snail_model_fpkm <- fpkm[classifications$snailModel == 'S']

snail_model_fpkm <- snail_model_fpkm[rowSums(snail_model_fpkm > 1 ) >= ncol(snail_model_fpkm),]
snail_model_fpkm_list <- list()

snail_model_fpkm_list$fpkm <- snail_model_fpkm
snail_model_fpkm_list$fpkm_scaled <- apply(snail_model_fpkm, 2, scale, center = T)
snail_model_fpkm_list$fpkm_scaled  <- as.data.frame(snail_model_fpkm_list$fpkm_scaled,row.names = rownames(snail_model_fpkm))

snail_model_fpkm_list$fpkm_log10 <- log10(snail_model_fpkm + 0.25 )
snail_model_fpkm_list$fpkm_scaled_log10 <- apply(snail_model_fpkm_list$fpkm_log10 , 2, scale, center = T)
snail_model_fpkm_list$fpkm_scaled_log10  <- as.data.frame(snail_model_fpkm_list$fpkm_scaled_log10,row.names = rownames(snail_model_fpkm))



snail_model_fpkm_som_list <- lapply(snail_model_fpkm_list, function(x)
                                { set.seed(7);  
                                  som(data=as.matrix(x), grid = somgrid(8,8, "rectangular"), keep.data =T)
                                }) 
snail_model_fpkm_somObj_list <- lapply(snail_model_fpkm_som_list, function(x){
                                somHelper$new(x)
                        })
fpkm_som_plots <- somPlotHelper$new(snail_model_fpkm_somObj_list)

fpkm_som_plots$plotAllCluster()
## Using cluster as id variables
## Using cluster as id variables

## Using cluster as id variables

## Using cluster as id variables

Note that the echo = FALSE parameter was added to the code chunk to prevent printing of the R code that generated the plot.